Modelling the temperature-induced blow-off and blow-on artefacts in filter-sorbent measurements of semivolatile substances

A common method of sampling semivolatile organic compounds (SVOCs) is to pull air through a filter to collect the particulate fraction followed by a gas-trapping sorbent (e.g. polyurethane foam or XAD resin) to collect the vapour fraction. This sampling method is prone to artefacts that may misrepresent the particle/gas distribution due to volatilisation from, and sorption to, the particles collected on the filter and the filter material itself. Such artefacts could be due to temperature variations as well as changes in atmospheric concentrations of sorbing species and in the mass of particles on the filter over the sampling period. Mathematical simulations of the changing equilibrium between particle and gas SVOCs during air sampling show that, in the absence of other artefacts, partition coefficients based on 24-h samples will match those in air at the end of the sampling period under warm conditions as well as cool conditions for the more volatile SVOCs. However, to reflect the average conditions of the sampling period, 24-h samples should be started near one of the two times that the average daily temperature occurs. If conditions are such that the equilibration needs of the collected particles cannot be met at all times during the sampling period, samples should be started at the morning occurrence of the average temperature. For the conditions in a mid-latitude urban area simulated in this study, 24-h samples started in early morning can yield temperature-induced blow-on and blow-off errors in estimated partition coefficients of up to 0.4 log Kp units.